RESUME. 91 



these except where they found an entrance at the edge. After making vain 

 attempts at one or two masked flowers, they usually hovered over others, 

 apparently inspecting them in the light of their previous experience. The 

 presence of a strange substance in the flower changed the relations so greatly 

 and offered an obstancle so foreign to the experience of the bee that it could 

 be solved in but few cases. Their behavior demonstrated the good memory 

 of bees for position in the flower, and their ignoring of the masked flowers, 

 both with and without experience of the cotton as an obstacle, indicated 

 their perception of form. 



Mutilation. — In more than three- fourths of the experiments mutilated 

 flowers were visited to a greater extent than the normal ones, the outstanding 

 exception being Monarda, to which the normal visits were five times more 

 numerous than those to the mutilated flowers. The advantage enjoyed 

 by the mutilated flowers as a whole was due chiefly to those in which both 

 stamens and styles were excised or all parts removed but the ovary and 

 nectaries. These changes not only permitted the odor of the nectar to 

 escape more freely, but also made the nectar itself more readily accessible, 

 with the consequence that the bees could work more flowers in the same time, 

 and were also led to return to them sooner. Increasing the attractive surface 

 of the perianth also augmented the number of visits; for example, splitting 

 the hood of Aconitum and turning the sepals back rendered such flowers 

 more than four times as attractive as the normal ones. On the contrary, 

 reducing the expanse of the corolla to a half diminished its attraction from 

 one-half to a tenth in practically all cases and demonstrated its paramount 

 importance even in the case of habituated bees. When both lips were 

 removed in Monarda, the number of visits was reduced to a fortieth of the 

 total for all mutilations, though the difficulty of landing obviously played 

 some part in the result. The removal of the lower lip alone decreased the 

 visits to a fourth of those to flowers with the upper lip alone removed and to 

 a tenth of the total visits to mutilations. 



The accounts of the behavior of the different species and of the individuals 

 of each species to the various mutilations naturally reveal similar specific 

 and individual differences to those found for changes of position, and these 

 are shown most clearly for Monarda in table 62. Clisodon paid only 1 visit 

 in 10 to the mutilated flowers in contrast to about 1 to 2 for Erynnis and 

 1 to 3 for Bombus juxtus. In contrast to the last, B. morrisoni made 1 visit 

 in 30 to mutilated ones, while B. occidentalis went to neither normal nor 

 mutilated Monarda, but only to flowers of competing species. There was 

 no striking difference between the general behavior of Bombus on the one 

 hand and the butterflies on the other. With respect to the type of mutila- 

 tion most visited, Erynnis, B. appositus, and B. juxtus went to each of the 

 5 types, Clisodon to 4, and Argynnis to but 3. Of the mutilated flowers 

 Clisodon gave the preference to those with the upper lip removed, as did 

 B. appositus; B. juxtus preferred those with the upper lip and stamens cut 

 away, while both Erynnis and Argynnis likewise went most often to the 

 flowers without an upper lip, the removal of the latter obviously interfering 

 least with attraction and access. 



The evidence of this table, as well as that from the mutilation experi- 

 ments as a whole, leaves no doubt that these pollinators were able to dis- 



